An important field of chemoinformatics consists in the pre-diction of molecule's properties, and within this field, graph kernels con-stitute a powerful framework thanks to their ability to combine a natural encoding of molecules by graphs, with classical statistical tools. Unfor-tunately some molecules encoded by a same graph and differing only by the three dimensional orientation of their atoms in space have different properties. Such molecules are called stereoisomers. These latter proper-ties can not be predicted by usual graph methods. In this report, ordered graphs are introduced in order to represent those molecules. Then the stereoisomerism property of each atom of a molecule is encoded by a lo-cal ordered subgraph. Finally a graph of interactions between those local subgraphs is constructed.